Bioinformatics Analysis To Identify Potential Biomarkers Of Esophageal Cancer And Prognostic Value

Objective: This study uses the sample data of esophageal cancer in the Cancer Genome Atlas database,and applies bioinformatics analysis methods to construct a prognostic risk scoring system based on CIMP grouping,mRNA,lncRNA,miRNA,and clinicopathological characteristics,which provides powerful prognosis and treatment decisions for esophageal cancer The theoretical basis.Methods: DNA methylation,mRNA,lncRNA,miRNA and clinicopathological data of esophageal cancer in the CGA database.(1)differential site analysis and K-mean clustering on DNA methylation data to obtain the CpG island methylation phenotype(CIMP)subgroup.(2)Difference analysis and enrichment analysis on RNA data,and construct ceRNA network and protein interaction network.Use Lasso regression,univariate and multivariate Cox regression to determine a set of RNAs related to the prognosis of esophageal cancer and establish a prognostic risk score for esophageal cancer.Use CIMP grouping,RNA prognostic risk score and clinicopathological characteristics to construct a nomogram model,and pass ROC The curve and decision curve evaluate the pros and cons of the nomogram model,and finally use resampling to verify the model internally.Results: A total of 46 DNA methylation sites were obtained through NA methylation data screening,and CIMP phenotype groups were obtained by clustering these sites,including CIMP-L,CIMP-M and CIMP-H.Survival analysis showed that the log-rank test of CIMP grouping was P0.001,CIMP-H had poor overall survival,which proved that CIMP grouping is an independent prognostic risk factor for esophageal cancer.Enrichment analysis for differentially expressed genes,including biological pathways,cellular component molecular functions and KEGG signaling pathway.Use 8 mRNA,8 miRNA and 25 lncRNA to construct a ceRNA network,and find 10 key genes through protein interaction network.After screening,a total of 12 RNAs related to prognosis were obtained,including 5 mRNAs(SLC26A9,COX6B2,OSM,RXFP3,RP13-672B3.2),4 lncRNAs(BLACAT1,CTD-2034I21.2,RP11-60A24.3),RP11-1123I8.1)and 3 miRNAs(hsa-mir-1269 a,hsa-mir-135 b,hsa-mir-935),respectively,to calculate the three prognostic risk scores.The nomogram model evaluation showed that the AUC of the combined model of CIMP grouping,RNA prognostic risk score,and clinicopathological characteristics was 0.84,showing good discrimination ability.The decision curve shows that the net clinical benefit of the model after increasing the RNA risk score is higher than that of the pure TNM staging model.The internal verification results show that the predicted value of the joint model and the actual value are in good agreement.Conclusion: This study found that CIMP is related to the deterioration of the overall survival of esophageal cancer,and the prognosis of patients with CIMP-H status is poor.It can be speculated that CIMP-related biomarkers may play a vital role in the occurrence of liver cancer.CIMP grouping can greatly promote the decision-making process of clinicians,and related genes may serve as biomarkers for the prognosis of esophageal cancer.The ceRNA network including these lncRNA,miRNA and mRNA differential expression profiles provides a new basis for further understanding the molecular mechanism of the occurrence and development of esophageal cancer.The protein interaction network screened out key genes to provide researchers with understanding of the response mechanism of esophageal cancer biosignals and energy metabolism,as well as understanding the functional connections between proteins.Finally,the prognostic risk score based on CIMP grouping and 12 RNA has good distinguishing ability,which is helpful for clinical treatment and prognostic decision-making of esophageal cancer,and provides more biomarker options for the prognosis and treatment of esophageal cancer.